Metal nanoparticles, nanowires, and carbon nanotubes* · Metal nanoparticles, nanowires, and carbon nanotubes 23 © 2000 IUPAC, Pure and Applied Chemistry 72, 21–33 4d states have

Pure Appl. Chem., Vol. 72, Nos. 1–2, pp. 21–33, 2000.© 2000 IUPAC

21

*Pure Appl. Chem. 72, 1–331 (2000). An issue of reviews and research papers based on lectures presented at the1st IUPAC Workshop on Advanced Materials (WAM1), Hong Kong, July 1999, on the theme of nanostructured systems.†Corresponding author: E-mail: [email protected]

Metal nanoparticles, nanowires, and carbonnanotubes*

C. N. R. Rao1,2,3†, G. U. Kulkarni1, A. Govindaraj3, B. C. Satishkumar1,and P. John Thomas1

1Chemistry and Physics of Materials Unit or 2CSIR Centre of Excellence inChemistry, Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur P.O.Bangalore 560 064, India; 3Solid State and Structural Chemistry Unit, IndianInstitute of Science, Bangalore 560 012, India

Abstract: The size-dependent metal to nonmetal transition in metal nanoparticles has beeninvestigated using photoelectron and tunneling spectroscopic techniques. Metal nanoparticlescapped by thiols are shown to organize into ordered 2D and 3D structures. Single-wallednanotubes and aligned carbon nanotube bundles have been synthesized by controlling thesize of metal nanoparticles produced in situ during the pyrolysis of precursors. Nanowires ofgold and other metals have been produced in the capillaries of the single-walled nanotubes.

INTRODUCTION

Electron transport in nanostructures, in particular in quantum dots and their arrays, is not only of academicinterest, but also holds promise in nanoelectronics [1–3]. Synthesis of metal nanoparticles and nanowiresof desired properties, therefore, assumes significance. Carbon nanotubes constitute yet another relatedclass of materials with several potential applications in electronics [4]. Of particular relevance in thiscontext, is the synthesis of aligned nanotube bundles.

In the last few years, metal nanoparticles have been investigated by various chemical and physicalmethods [1]. Size-dependent electronic structure of the nanoparticles is the focus of these studies. It isgenerally known that the conduction band present in a bulk metal will be absent in a nanoparticle, andinstead there would be discrete states at the band edge. Electrons may therefore undergo quantum-confinement in very small metal particle, showing properties of a quantum dot [2]. Of particular interestare assemblies of colloidal particles [5], wherein metal nanoparticles, suitably covered by organic mol-ecules such as thiols, get self-organized in crystalline arrays of one, two, or three dimensions. Electrontransport in such mesostructures is equally interesting. In this article, we briefly present the size-depen-dent electronic properties of the metal particles we investigated and provide some highlights of theirassemblies in two and three dimensions. We also present the synthesis of bundles of aligned carbonnanotubes employing a pyrolysis method where magnetic metal nanoparticles produced in situ play acrucial role. Metal nanowires produced in the capillaries of single-walled nanotubes are also discussed.

ELECTRONIC STRUCTURE OF METAL NANOPARTICLES

The electronic structure of a metal particle critically depends on its size. For small particles, the electronicstates are not continuous, but discrete, due to confinement of the electron wavefunction. The averagespacing of successive quantum levels, d, known as the Kubo gap, is given by d = 4Ef /3n, where Ef is the

22 C. N. R. RAO et al.

© 2000 IUPAC, Pure and Applied Chemistry 72, 21–33

Fermi energy of the bulk metal and n is the number of valence electrons in the nanoparticle (usuallytaken as its nuclearity). Thus, for a Ag nanoparticle of 3-nm diameter containing ~103 atoms, the valueof d would be 5–10 meV. Since at room temperature, kT @ 25 meV, the 3-nm particle would be metallic(kT > d). At low temperatures, the level spacings especially in small particles, may become comparableto kT, rendering them nonmetallic [2]. Because of the presence of the Kubo gap, properties such aselectrical conductivity and magnetic susceptibility exhibit quantum size effects. Discreteness of energylevels also brings about changes in the spectral features, especially those related to the valence band.

High-energy spectroscopies provide a direct access to the electronic structure of metal nanoparticles.In X-ray photoelectron spectroscopy (XPS), the kinetic energy of the photoelectron ejected from anatomic core-level is measured to determine the binding energy of that level. Ultraviolet photoelectronspectroscopy (UPS) and bremmstrahlung isochromat spectroscopy (BIS), on the other hand, provideinformation on the occupied and unoccupied levels respectively near Ef. In the last few years, severalexperiments have been carried out on nanoparticles in the desired size range obtained by the depositionof metals on amorphized graphite and other substrates, using electron spectroscopic techniques [6,7].The particles were obtained by the resistive evaporation of the corresponding metal in vacuum. Thevariation in the core-level binding energy of the nanoparticle with respect to the bulk metal value hasbeen measured as a function of the coverage or particle size for metals such as Au, Ag, Pd, Ni, and Cu.An important result from these experiments is that as the particle size decreases, the binding energyincreases sharply by as much as 1.2 eV as in the case of Pd (Fig. 1). The variation is negligible at largecoverages or particle size since the binding energies would be close to those of the bulk metals. Theincrease in the core-level binding energy in small particles occurs due to the poor screening of the core-hole and is a manifestation of the size-induced metal–insulator transition in nanoparticles.

The electronic band structure of metal nanoparticles near the Fermi level as revealed by UPS andBIS techniques is of considerable interest [6–8]. In Fig. 2 we show typical spectra of Pd nanoparticles.For large particles (~5 nm diameter), the UP spectrum is similar to that of the bulk metal, where aconsiderable intensity is present at Ef due to the Pd(4d) states arising from the metallic nature of thenanoparticles. For small particles (<2 nm diameter), however, the intensity of the 4d states at Ef de-creases rapidly, accompanied by a shift in the intensity maximum to higher binding energies and anarrowing of the 4d-related spectral features. The BI spectra of small nanoparticles show that the empty

Fig. 1 Variation of the shifts in the core-level binding energies (relative to the bulk metal value) of Pd (a) and Au(b) clusters, with the average diameter of the cluster. The diameters were obtained from HREM and STMimages. In the case of Au clusters, data for a colloidal particle and a Au55 compound are also shown.

Metal nanoparticles, nanowires, and carbon nanotubes 23


4d states have negligible intensity at Ef. With the increasing particle size, new states emerge closer to Efresembling the spectrum of bulk metal. Thus, both UP and BI spectra establish the emergence of newstates at Ef with the increase in the particle size, consistent with the occurrence of an insulator-to-metaltransition. Similarly, the BI spectra of Ag nanoparticles show a distinct feature that moves towards theEf with increase in size, accompanied by an increase in the intensity of the 5s band. All these electronspectroscopic measurements indicate that a gap manifests itself around a nanoparticle diameter of 1–2nm possessing 300 ± 100 atoms. Recent photoelectron spectroscopic measurements [8] on mass-se-lected Hgn nanoparticles (n = 3 to 250) reveal that the HOMO-LUMO (s-p) gap decreases graduallyfrom ~3.5 eV for n = 3 to ~ 0.2 eV for n = 250. The bandgap closure is predicted at n ~ 400. In metalcolloids, surface plasmon excitations impart characteristic colors to the metal sols, the wine-red color ofthe gold sols being well-known. We found that colloidal particles of 4.2 and 2.1 nm diameters exhibit adistinct band around ~525 nm [9], the intensity of which increases with size. The intensity of thisfeature becomes rather small in the case of the 1-nm-diameter particles (£200 atoms).

Direct information on the nonmetallic gap states in nanoparticles is obtained by scanning tunnel-ing spectroscopy (STS). This technique provides the desired sensitivity and spatial resolution, making itpossible to carry out tunneling spectroscopic measurements on individual particles. A systematic STSstudy of Pd, Ag, Cd, and Au nanoparticles of varying sizes deposited on a HOPG substrate has beencarried out recently under ultra-high vacuum conditions, after having characterized the nanoparticlesby XPS and STM [10]. The overall slope of the I-V curve near zero-bias, which is related to the densityof states (DOS) near Ef, increases linearly up to particle volume of ~4 nm3 (~2 nm diameter) andbecomes constant for bigger particles as shown in Fig. 3a. Accordingly, the DOS in particles with adiameter >2 nm (with ~300 atoms or more) resembles that of the bulk. Derivative spectra of biggerparticles were featureless, while those of the small particles (<1 nm) showed well-defined peaks oneither side of zero-bias due to the presence of a gap. Ignoring gap values below 25 meV (~kT), it is seenthat small particles of £1 nm diameter are nonmetallic (Fig. 3b). This observation is consistent with the

Fig. 2 Electronic structure of small Pd clusters near the Fermi level. (a) He II UP spectra show a decrease in theintensity of the occupied 4d band with the decreasing cluster size. This is accompanied by a shift in the intensitymaximum to higher binding energies and a narrowing of the spectral features. (b) BI spectra show similar effectsin the empty 4d states above the Fermi level. IPd/IC values (ratios of XPS core-level intensities of Pd and thegraphite substrates) are used as a measure of the coverage or the cluster size. The approximate cluster diameter isindicated in extreme cases.



absence of the plasmon band in nanoparticles of 1 nm diameter. From the various studies discussedhitherto, it appears that the size-induced metal-insulator transition in metal nanoparticles occurs in therange of 1–2 nm dia or 300 ± 100 atoms. Theoretical investigations of the electronic structure of metalnanoparticles also throw light on the size-induced changes in the electronic structure. Rosenblit andJortner [11] calculated the electronic structure of a model metal cluster and predicted electron localiza-tion to occur in a cluster of diameter ~0.6 nm.

Bimetallic nanoparticles

Bimetallic or alloy colloids have been made by the chemical reduction of the appropriate salt mixturesin solution phase. Thus, Ag–Pd and Cu–Pd colloids of varying compositions have been prepared byalcohol reduction of AgNO3+PdOx and CuOx+PdOx mixtures [12]. Ag–Pt colloids have been preparedby the NaBH4 reduction of oxalate precursors and Cu–Pd colloids by the thermal decomposition of

Fig. 3 Scanning tunnelingspectroscopy of individual metalclusters: (a) Conductance (obtainedfrom the normalized slope of the I-Vcurves) as a function of clustervolume. Above a critical volume~4 nm3, the slope becomes size-independent. (b) The conductiongap observed in small clusters ofAu, Pd, Cd, and Ag as a function ofcluster volume. The volume wascalculated from the cluster diameterobtained from STM.



acetate mixtures in high-boiling solvents. Several bare bimetallic nanoparticles have been deposited onsolid substrates by thermal evaporation in vacuum [13]. Examples of such alloy nanoparticles areNi–Cu, Ag–Au, Ni–Pd and Cu–Pd. Measurements of the core-level binding energies of the componentmetals in the alloy particles show that the shifts in binding energy (relative to those of bulk metals), DE,comprise additive contributions from alloying and the particle size. That is, the observed DE is given by,DE = DEalloy + DEsize. At large particle size, only the alloying effect, DEalloy, is found as in bulk alloys.Small alloy particles exhibit the size effect, in that the DEsize shows a marked increase, with decrease inparticle size. The effect due to particle size is shown in Fig. 4 in the case of Ni–Cu alloy nanoparticles.

NANOPARTICLE ASSEMBLIES

The quest for nanoscale architecture has demanded newer synthetic methodologies for forming andorganizing metal particles. In this context, the use of self-assembling surfactants is of significance.Thus, metal nanoparticles capped with thiols, silanes, phospholipids, and phosphines can be prepared insolid form. We shall examine the recent results obtained with 2D and 3D assemblies of metal nanoparticles.

Two-dimensional arrays

Schiffrin and co-workers [14] first prepared gold organosols using alkane thiols as surfactants, by phase-transferring gold ions and carrying out reduction in the presence of the thiols. A novel method of thiol-

Fig. 4 Coverage or cluster sizedependence of the shift in the 2p3/2binding energy of (a) Cu and (b) Ni inCu–Ni bimetallic clusters.



derivatizing hydrosols of various metals has been developed recently [15]. The procedure involvesmixing a hydrosol containing metal particles of the desired size distribution with a toluene solution ofan alkane thiol (butanethiol or higher members). The immiscible liquid layers thus obtained is stirredvigorously with HCl or NaBH4 when the metal particles in the bottom aqueous layer gush to the upperhydrocarbon layer containing the thiol and get thiol-derivatized in this process. The completion of thederivatization is marked by a vivid interchange of the colors from the aqueous layer to the hydrocarbonlayer. The advantage of this method is that well-characterized metal particles can be easily thiol-derivatizedin a nonaqueous medium. Gold, silver, and platinum particles of different sizes have been thiolized bythis procedure.

Thiolized metal nanoparticles from a nonaqueous medium are readily assembled into 2D arrayson a suitable solid substrate such as HOPG. As an example, the TEM image of a nanocrystalline arrayof gold particles with a mean diameter of 4.2 nm is shown in Fig. 5. The metal particles form close-packed structures extending over tens of nms with a regular spacing of ~1 nm between them [9]. X-raydiffraction pattern of this array exhibits a low-angle peak corresponding to a d spacing of 5.0 nm(Fig. 5). The distance between the particles deduced from the d-spacings is somewhat smaller than thatexpected from the dimensions of the metal particles and the thiol, suggesting some overlap of the alkane

Fig. 5 2D array of thiol-derivatized Au particles of 4.2 nm mean diameter. Histograms indicating particle sizedistribution is given. XRD pattern from this array is also shown.



chains of the thiols on the neighboring particles. Ordered 2D lattices containing thiolized Au particlesof two different sizes have been reported recently by Kiely et al. [16]. A careful study in this laboratoryof 2D lattices formed by Pd nanoparticles of different sizes by employing alkane thiols of differentlengths (butane to hexadecanethiol) has revealed some interesting aspects of particle assemblies. Or-dered 2D lattices are formed with Pd particles in the diameter range of 2 to 6 nm with octanethiol anddodecanethiol. Satisfactory 2D lattices are generally observed when the ratio of the particle diameter(d) and the thiol chain length (l) is between 1 and 4, the most ordered lattices being favored when thed/l ratio is between 2 and 3.

It is of considerable value if one can prepare 2D lattices of metal quantum dots containing fixednumber of atoms. Langmuir–Blodgett films containing Au55 nanoparticles have been obtained [17].More interestingly, starting with cluster compounds of fixed nuclearity, it is possible to prepare the 2Dlattices of thiolized Pd particles of uniform size. In Fig. 6, we show a TEM image of the 2D array of Pdparticles, each containing around 1400 atoms.

Thiol-capped metal nanoparticles provide an excellent means of studying monoelectron conduc-tion in metal quantum dots, giving rise to the Coulomb staircase phenomenon. In Fig. 7, we show howthe charging energy of Au particles varies with the size. The Coulomb gap increases with the decreasein particle size as expected. The electronic structure of 2D lattices of metal quantum dots is of interest.A 2D lattice consisting of metal nanoparticles separated by a distance using an organic spacer serves asa model system to study the Mott–Hubbard metal–insulator transition wherein a decrease in the dis-tance between the quantum dots (or increase in pressure) closes the Coulomb gap. By varying theinterparticle distance continually in a 2D lattice of Ag particles (3 nm diameter) in a Langmuir trough,Collier et al. [18] have indeed found the Coulomb gap to vanish at a critical distance. We are examiningmetal particles linked by conjugated thiols and dithiols of varying lengths, in comparison with those ofsaturated thiols.

Three-dimensional superlattices

Multilayer assemblies using monothiols are generally fragile. Multilayer deposition of particle arrays isbest achieved by the sequential adsorption of dithiol molecules and metal nanoparticles of the desiredsize, by dipping the substrate into the respective solutions with intermediate steps involving washingwith toluene and drying [19]. Using this procedure, several monometal, bimetal, and metal-semiconductorsuperlattices have been prepared. The procedure is shown schematically in Fig. 8. As many as five

Fig. 6 TEM image of 2D array of thiol-derivatized Pd clusters of uniform sizeand nuclearity. Inset shows lattice fringesof individual clusters corresponding to anuclearity of around 1400.



Fig. 7 Charging energies of PVP-covered gold nanoparticle of varyingsizes. A coulomb staircase obtainedwith a 4-nm-diameter particle is shownin the inset.

Fig. 8 Schematic drawing depictingthe layer-by-layer deposition of Ptparticles onto a Au substrate, thelayers being separated by dithiolmolecules. Also shown is theformation of a heterostructureconsisting of alternate layers ofsemiconductor and metal particles.



depositions of nanocrystalline arrays could be accomplished by this method. After each deposition, thestructure was characterized by STM and X-ray diffraction, as well as by XPS. STM images showed thepresence of regular arrays of nanoparticles extending over 300 nm, corresponding to the size of a typicalflat terrace on the substrate. The images also revealed a nearly regular spacing of 2 nm between thenanoparticles. XRD patterns recorded after successive depositions of layers exhibited low-anglereflections with the d-spacings reflecting the particle diameter and the inter-particle distance. Theintensities of Pt(4f), C(1s) and S(2p) core-levels increased while Au(4f), due to the substrate, decreasedwith successive depositions. A plot of the metal coverage versus the number of depositions gave a slopeclose to unity until the third deposition and increased thereafter, suggesting that layer-by-layer depositionof nanoparticles had been accomplished.

ROLE OF METAL NANOPARTICLES IN THE FORMATION OF CARBON NANOTUBES

Depending on the size of the metal nanoparticles, pyrolysis of hydrocarbons produces single-wallednanotubes (SWNT), multiwalled nanotubes (MWNT), or metal-filled onions or graphene-covered metallicparticles. We illustrate this in Fig. 9. Organized metal nanoparticles are used to produce aligned bundlesof nanotubes by the pyrolysis of hydrocarbons. Small particles with diameters in the 1–2 nm rangecatalyze the formation of SWNTs while bigger particles of ~10–50 nm give rise to MWNTs. Bycontrolling the partial pressure of the organometallic and hydrocarbon precursors, one can preferentiallyobtain SWNTs [20].

Fig. 9 Schematic illustration of therole of metal particles in theformation of carbon nanotubes bythe pyrolysis method. Smallparticles (~1 nm) give rise toSWNT, bigger particles (~10 nm)produce MWNT while muchlarger particles (>50 nm) getencapsulated in graphite layers.



Aligned nanotube bundles have been obtained previously by chemical vapor deposition on laser-patterned substrates [21] or in porous solids [22,23]. We have produced aligned nanotubes by the py-rolysis of mixtures of organometallic precursors and hydrocarbons [24]. The advantage of this methodis that aligned tubes may be produced in one step, without the use of a substrate. In Fig. 10, we show theSEM images of aligned nanotubes obtained by the pyrolysis of ferrocene with methane, acetylene, andbutane respectively at 1373 K. The average length of the nanotubes was generally around 60 µm withmethane and acetylene. Although the nanotubes are aligned in the case of methane (Fig. 10a), thepacking density is not high. With acetylene, however, more compact nanotube bundles are obtained, asseen from Fig. 10b. A negligible proportion of graphene-covered metal nanoparticles were presentalong with the nanotubes when methane or acetylene were used as the hydrocarbon sources. Whenbutane was used (Fig. 10c), there was a greater proportion of extraneous carbon deposits mainly con-sisting of Fe particles encapsulated inside graphitic sheaths. The average length of the nanotubes wasalso shorter.

In order to study the nature of the metal nanoparticles involved in the formation of the nanotubebundles, we have analyzed the samples using TEM. As an example, a TEM image of the nanotubesobtained by the pyrolysis of ferrocene–acetylene mixture is shown in Fig. 11. Iron nanoparticles stick-ing to the end portions of the nanotubes can be clearly seen. The particle diameters are in the range of2–13 nm. Magnetization measurements carried out on this sample showed the saturation magnetizationto be ~20 emu/g, which is considerably smaller than the bulk value of 222 emu/g. It is noteworthy that

Fig. 10 SEM images (viewed perpendicular to thenanotube axis) of aligned nanotubes obtained fromferrocene–hydrocarbon mixtures; (a) methane,(b) acetylene, and (c) butane.



Fig. 11 TEM image of a part of an aligned nanotube bundle obtained from the pyrolysis of acetylene–ferrocenemixture, showing Fe nanoparticles along with the nanotubes.

Fig. 12 (a) TEM image of Au nanowires inside SWNTs obtained by the sealed tube reaction. Inset shows aSAED pattern of the nanowires. (b) HREM image showing the single crystalline nature of some of the Aunanowires.



the Fe particles involved in the formation of carbon nanotubes are sufficiently large to render themferromagnetic. We therefore suggest that the ferromagnetism of the Fe nanoparticles may be respon-sible for the alignment of the nanotubes produced by pyrolysis of precursors.

METAL NANOWIRES

Metal nanowires encapsulated in carbon nanotubes have been obtained by treating the SWNTs withmetal salts at melting temperatures in vacuum-sealed quartz tubes followed by hydrogen reduction[25]. In Fig. 12a, we show a TEM image which reveals the presence of a large number of nanotubesfilled extensively with gold. The length of the gold wires so obtained is in the range of 15–70 nm, andthe diameters in the range of 1.0–1.4 nm. Selected area electron diffraction (SAED) patterns taken fromthe region of nanowires (inset of Fig. 12a) shows diffuse rings due to small particles. The diffractionrings correspond to the (111) and (002) reflections of gold. In some of the nanowires, we have found themetal to be single crystalline as shown in the HREM image in Fig. 12b. This image reveals the resolvedlattice of gold with a spacing of ~0.23 nm, corresponding to the (111) planes. Polycrystalline Auaggregates could be transformed into single crystalline wires by suitable annealing. UV-vis spectra ofdispersions of the gold nanowires in ethanol show transverse plasmon absorption band at ~520 nm andlongitudinal bands at 670, 860 and 1150 nm. These bands are an indication of the quantum confinementexperienced by the gold nanowires, and also they imply different aspect ratios (3–6) of the wires [26].Employing similar methods, nanowires of Ag, Pt, and Pd have also been obtained in the capillaries ofSWNTs.

REFERENCES

1. J. H. Fendler (ed.). Nanoparticles and Nanostructured Films, Wiley-VCH, Weinheim (1998).2. P. P. Edwards, R. L. Johnston, C. N. R. Rao. In Metal Clusters in Chemistry, P. Braunstein, G. Oro,

P. R. Raithby (eds.), Wiley-VCH, Weinheim (1999).3. G. Markovich, C. P. Collier, S. E. Henrichs, F. Remacle, R. D. Levine, J. R. Heath. Acc. Chem.

Res. 32, 397 (1999).4. S. J. Tans, A. R. M. Verschueren, C. Dekker. Nature 393, 49 (1998).5. C. P. Collier, T. Vossmeyer, J. R. Heath. Annu. Rev. Phys. Chem. 49, 371 (1998).6. H. N. Aiyer, V. Vijayakrishnan, G. N. Subbanna, C. N. R. Rao. Surf. Sci. 313, 392 (1994).7. V. Vijayakrishnan, A. Chainani, D. D. Sarma, C. N. R. Rao. J. Phys. Chem. 96, 8679 (1992).8. R. Busani, M. Follera, O. Chesnovsky. Phys. Rev. Lett. 81, 3836 (1998).9. K.V. Sarathy, G. Raina, R. T. Yadav, G. U. Kulkarni, C. N. R. Rao. J. Phys. Chem. B 101, 9876

(1997).10. C. P. Vinod, G. U. Kulkarni, C. N. R. Rao. Chem. Phys. Lett. 289, 329 (1998).11. M. Rosenblit and J. Jortner. J. Phys. Chem. 98, 9365 (1994).12. H. N. Vasan and C. N. R. Rao. J. Mater. Chem. 5, 1755 (1995) and references therein.13. K. R. Hari Kumar, S. Ghosh, C. N. R. Rao. J. Phys. Chem. 101, 536 (1997).14. M. Brust, M. Walker, D. Bethell, D. J. Schiffrin, R. Whyman. J. Chem. Soc. Chem. Commun. 801

(1994).15. K. V. Sarathy, G. U. Kulkarni, C. N. R. Rao. J. Chem. Soc. Chem. Commun. 537 (1997).16. C. J. Kiely, J. Fink, M. Brust, D. Bethell, D. J. Schiffrin. Nature 396, 444 (1998).17. L. F. Chi, S. Rakers, J. Drechesler, M. Hartig, H. Fuchs, G. Schmidt. Thin Solid Films 329, 520

(1998).



18. C. P. Collier, R. J. Saykally, J. J. Shiang, S. E. Henrichs, J. R. Heath. Science 277, 1978 (1997).19. K. V. Sarathy, P. J. Thomas, G. U. Kulkarni, C. N. R. Rao. J. Phys. Chem. B 103, 399 (1999).20. C. N. R. Rao, A. Govindaraj, R. Sen, B. C. Satishkumar. Mat. Res. Innovat. 2, 128 (1998).21. M. Terrones, N. Grobert, J. P. Zhang, H. Torrenes, J. Olivares, H. K. Hsu, A. K. Cheetam,

H. W. Kroto, D. R. M. Walton. Chem. Phys. Lett. 285, 299 (1998).22. W. Z. Li, S. S. Xe, L. X. Qian, B. H. Chang, B. S. Zou, W. Y. Zhou, R. A. Zhao, G. Wang. Science

274, 1701 (1996).23. G. Che, B. B. Laxmi, C. R. Martin, E. R. Fischer, R. S. Ruoff. Chem. Mater. 10, 260 (1998).24. B. C. Satish Kumar, A. Govindaraj, C. N. R. Rao. Chem. Phys. Lett. 307, 158 (1999).25. A. Govindaraj, B. C. Satish Kumar, M. Nath, C. N. R. Rao. Chem. Mat. 12, 202 (2000).26. S. Link, M. B. Mohammed, M. A. El-Sayed. J. Phys. Chem. B 103, 3073 (1999).

Documents

Metal nanoparticles, nanowires, and carbon nanotubes* · Metal nanoparticles, nanowires, and carbon nanotubes 23 © 2000 IUPAC, Pure and Applied Chemistry 72, 21–33 4d states have